Application of Pattern Recognition Techniques for the Analysis of Thin Blood Smear Images

Analysis of microscopic medical images is an important interdisciplinary problem involving both physicians and computer scientists. One of the important and active areas of research is the problem of counting blood cells (CBC) [12] which is used as screening test to check such disorders as infections, allergies, problems with clotting, and it helps diagnosing and managing a large number of diseases. In practice a panel of tests is carried out that examine different blood components such as counting white blood cells (WBC) [12-Ch.159, 1], white blood cells differential, counting red blood cells (RBC) [12-Ch.159], checking for signs of disease and the counting the number of infected cells. In normal human blood, there are 4,000,000-6,000,000, 4,000-11,000, 150,000-450,000 per microliter of RBC, WBC, and normal platelet counts, respectively, with platelets usually present in complexes rather than singularly [12]. But instead of the special case of spontaneous bleeding, platelet counts are rarely requested in a CBC, so in this work we will focus on RBC and WBC counts. A manual diagnosis would search for abnormalities in the blood cells and particles while performing a CBC. Complications may arise from the large number of hematic pathologies [31] including the large number of WBC sub-types [22], which makes the analysis prone to human error. This process can be automated by computerized techniques which are more reliable and economical. Therefore there is always a need for the development of systems to provide assistance to hematologists and to relieve the physician of drudgery or repetitive work. Our goal is to develop and validate an image processing and pattern recognition system to quantify and detect microscopic particles on thin blood smear slides to enhance automated system to characterize blood health status of patient. In essence we seek to determine a fast, accurate mechanism for segmentation and gather information about distribution of microscopic particles which may help to diagnose the degree of any abnormalities during clinical analysis. Automatic detection of pathologies from histopathological images is currently very active and important area of research.